Generative Artificial Intelligence (AI) is finding its pathway into diverse industries, manifesting its potential in unique ways. A fundamental example of this is Google’s Bard – an AI developed by Google that has made headlines recently. In India, the Ministry of Electronics and IT’s (MeitY) is implementing the power of AI through BHASHINI, which is building a ChatGPT-Powered WhatsApp Chatbot to aid Indian farmers.
BHASHINI, short for BHASHa INterface for India, is a language translation platform led by Artificial Intelligence in India. The development and release of the WhatsApp chatbot are currently undergoing due to the limitation of support for local languages, as ChatGPT presently relies on English input.
Intended Features of the ChatGPT-Powered WhatsApp Chatbot
The proposed WhatsApp chatbot will allow users to provide queries using voice notes and receive voice-based responses generated by ChatGPT. This feature is specifically designed keeping in mind India’s rural and agrarian population, which heavily depends on government schemes and subsidies.
Understanding that potential users speak diverse languages, the necessity to build a language model that can successfully identify and comprehend them has been acknowledged. This would offer immense help to Indian farmers who might not be adept at typing on smartphones.
The designed chatbot commits to supporting 12 languages including English, Hindi, Tamil, Telugu, Marathi, Bengali, Kannada, Odia, and Assamese. To assist users who don’t understand English, the Government’s Bhasha Daan Initiative will be utilised which crowdsources language inputs for various Indian languages as part of Project BHASHINI. This initiative encourages citizens to contribute to building an open repository of data to digitally enrich their own language.
Concerns Related to Generative AI Models
Despite the promising features, there are concerns associated with generative AI models like ChatGPT and Google’s Bard. There have been instances where these models have provided inaccurate information. A recent example includes an error made by Bard about the James Webb Space Telescope which led to a USD 100 billion drop in the company’s shares.
Currently, the WhatsApp chatbot is in its testing phase and can only respond to simple queries about government schemes due to the limitations of ChatGPT — it does not have access to real-time internet data. It was trained on a dataset encompassing information up to 2021.
However, this may change soon with Microsoft’s announcement about a new version of Bing, powered by an upgraded version of the same AI technology behind ChatGPT, GPT 3.5, known as the “Prometheus Model”.
About ChatGPT and Its Applications
ChatGPT represents a variant of the Generative Pre-trained Transformer (GPT), a large-scale, neural network-based language model developed by OpenAI. These models are trained on extensive text data to generate human-like text, capable of answering questions, providing explanations, and participating in conversations.
ChatGPT’s applications extend to various realms such as digital marketing, online content creation, customer service queries, and even debugging code. The bot imitates human speaking styles and can respond to a vast range of questions, making it a suitable candidate for writing basic emails, CVs, college essays, and even coding tasks.
This innovative technology holds the potential to revolutionize the way humans interact with AI, paving the way for a future where machines can effectively assist in a variety of tasks, reducing human workload and enhancing efficiency.